AI Summary & Key Takeaways
AI InsightResults: Overall, 90.1% were willing to use a secure institutional AI system compliant with national data protection requirements, whereas fewer than 15% would enter directly identifying information such as patient names or file numbers.
📌 3 Key Takeaways
Abstract
Aims: Artificial Intelligence (AI) is increasingly used in healthcare, but its routine use by pediatricians remains unclear. This study examined the clinical and academic tasks for which pediatricians use AI, their willingness to use secure institutional systems, and the patient information they would enter. These dimensions were examined together to distinguish general acceptance of AI from task-specific use and willingness to share different types of patient information. Methods: This cross-sectional web-based survey included 302 pediatricians in Turkiye in June-July 2026. A 25-item questionnaire assessed task-specific AI use, ethical and professional concerns, institutional preparedness, willingness to use secure systems, and patient data-sharing preferences. Content validity was evaluated by 10 experts using a modified Lawshe approach. Data were summarized descriptively; Kruskal-Wallis, Mann-Whitney U, and Spearman correlation analyses were performed. Demographic and professional analyses were exploratory. Results: Overall, 90.1% were willing to use a secure institutional AI system compliant with national data protection requirements, whereas fewer than 15% would enter directly identifying information such as patient names or file numbers. Clinical use was most common for medical information seeking (61.3%) and least common for preparing patient or family education materials (22.8%); 12.6% reported no clinical use. In academic work, literature searching and summarization was most frequent (68.9%), while project or grant applications were least frequent (7.6%); 11.6% reported no academic use. Older age was weakly associated with lower regular AI use (ρ=-0.192, p<0.001). Conclusion: Pediatricians use AI-supported language models across clinical and academic tasks and are receptive to secure institutional systems. However, willingness to share directly identifying patient data remains limited. Implementation should prioritize data protection, clinician oversight, and practical training.